2

I’m inquiring about the overlay mode, in one of its applications: emboss sharpening.

Particularly, a layer above can be overlayed to the layer beneath when it is applied with a suitable emboss filter to create a sharpening effect. I got the formula from the gimp docs. But, I couldn’t make sense out of it intuitively as with other simple layer modes.

What is the theory behind the emboss filter technique that sharpens the image?

Improve this question

1 Answer 1

Reset to default
0

The standard color mode used by images before the advent of hdri is a bit limited. It only has colors in range of 0-1. But obviously colors and especially signal processing is not limited to a range of 0-1. Especially negative colors would be nice to have.

The overlay mode sidesteps this limitation. By setting 0.5 color as neutral, and anything below is subtracted (or that is what it emulates, not what it does) to the signal and adds the signal above 0.5. Since emboss is effect that wants to do adding a subtracting on both sides it is aimed for overlay and produces a grayish image with darker shadows and lighter highlights.

Now emboss is part of a family of a edge detection/highpass filters. Now it turns out that if you boost edges you in fact get apparent sharpening, much like unsharp mask. The image is not really sharper, the local contrast is higher. Which is much the same for less than 1 pixel sharpening, and since the human eyes that actually do similar kind of sharpening on the retina it will interpret the signal sharper.*

* But yes we can also really sharpen images since in the era of Hubble myopia, but that does increase the noise...

Improve this answer
3
  • do scientists develop such techniques and designers just apply them? What does it take a designer to develop such techniques for image manipulation?(I suppose a solid understanding of the signal processing or mathematics perhaps would be essential?)
    – mathmaniage
    Jul 26, 2017 at 12:09
  • Well, it certainly helps to know what is happening in order to define how two pixels should be combined into a result. In this case, en.wikipedia.org/wiki/Blend_modes can be helpful to get an idea.
    – Michael Schumacher
    Jul 26, 2017 at 12:43
  • @BeshalJaenal Its a pretty old technique from photographers, but not all designers are image manipulation operators. Some have a very high education.
    – joojaa
    Jul 26, 2017 at 17:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.